BLM Juniper Mastication QA/QC
Timothy Littmann
2025-07-15
Brief Overview and Objective
Late season 2024 and 2025, 25 m transects were randomly placed within the treatment areas where Juniper mastication had occurred. Some of the transects were in treated areas and some outside (controls). All 2024 samplings took place pre-treatment. The sampling process consisted of 2 types of woody plant measurements (course woody debris and tree/shrub coverage) and 2 types of herbaceous measurements (plant composition and productivity). The purpose of this QA/QC is to evaluate and ensure procedures and documentation of sampling periods had been executed properly in support of the desired objective and accuracy of data. Protocols, notes, hard copies and methods have been reviewed as this document was being crafted.
Sampling conducted:
• 26 sampling locations were visited and recorded (2024) - June 24, 25/July 1
• 22 sampling locations were visited and recorded (2025) - June 9, 10, 11, 26
• For each location 1 transect 25m; 5 quadrats @ 5 m intervals
Data recorded
(note both 2024 and 2025 are compiled together into a single Excel file with each treatment type assigned its own sheet):
File path: OneDrive – USDA > Juniper mastication > BLM > data > CedarCreekMastication
• Course woody debris: a total of 47 observations from 7 variables had been recorded and entered. Observations consisted of numbers of small, medium and large size woody debris respectively. Larger than 8 cm in size were recorded by exact measurements.
• Tree/shrub coverage: a total of 589 observations from 5 variables had been recorded and entered. Observations consisted of “HeightHits” along a 1 x 25 meter transect, where the protocol dictated that hits under 1 m in height were measured in cm and over 1 meter in height was recorded by a start and end point along the transect (cm).
• Plant composition: a total of 2540 observations from 6 variables had been recorded and entered. Observations consisted of ground, structural and aerial measurements within quadrats using the Daubenmire method for estimating the ranges of percent coverage. The midpoints for each percentage range were set as the values to be recorded. Plant codes were applied for each species observed.
• Productivity: a total of 241 observations from 5 variables had been recorded and entered. Herbaceous vegetation within each quadrat had been clipped to within 3 cm of the surface (excluding woody material and other debris). 5 of the bags used for the sampling were tared empty and averaged. Samples were weighed and recorded.
• All hardcopies have been scanned and filed away.
Woody Debris
As noted earlier, 47 observations from 7 variables are recorded. No missing or duplicated entries detected.
No further action taken.
## # A tibble: 1 × 7
## missing_year missing_date missing_transect missing_small missing_medium
## <int> <int> <int> <int> <int>
## 1 0 0 0 0 0
## # ℹ 2 more variables: missing_large <int>, missing_Above8 <int>## # A tibble: 0 × 7
## # Groups: transect, year [0]
## # ℹ 7 variables: year <dbl>, date <dttm>, transect <chr>, small <dbl>,
## # medium <dbl>, large <dbl>, Above8 <chr>## # A tibble: 1 × 1
## Duplicates
## <int>
## 1 0Below are visual representations for the distribution of values of various sized woody debris comparing 2024 and 2025.
Noted Outliers shown below have been verified. Due to the negative skew from many zero entries, both the data and hard copies were examined to ensure accuracy.
## Outliers for small (2024): 90 70 7 10
## Outliers for small (2025): 77 112 59 70
## Outliers for medium (2024): 21 36 0 1
## Outliers for medium (2025): 75 59 76 64 111
## Outliers for large (2024): 15 7 9
## Outliers for large (2025): 12 24 16## Outliers for Above8 (2024): 17 21 21 14 20
## Outliers for Above8 (2025): 10 10 12Productivity
241 observations from 5 variables had been recorded and entered.
Upon examination, no entries were missing; However, duplicate entries were detected.
## # A tibble: 1 × 5
## Missing_Year Missing_Transect Missing_Quadrat Missing_GrossMass Duplicates
## <int> <int> <int> <int> <int>
## 1 0 0 0 0 3## # A tibble: 6 × 5
## # Groups: year, transect, quadrat [3]
## year transect quadrat GrossMass BagType
## <dbl> <chr> <dbl> <dbl> <chr>
## 1 2024 47 5 11.7 B
## 2 2024 47 5 10.2 B
## 3 2025 101B 4 34.9 L
## 4 2025 101B 4 30.6 L
## 5 2025 106B 3 30.3 B
## 6 2025 106B 3 30.4 BAction taken: Cross checked hardcopies and reweighed samples in question. Provided script to remove the incorrect entries. (see script below)
Clipping %>%
distinct(year, transect, quadrat, .keep_all = TRUE)Noted Outliers shown below have been verified. Both the data and hard copies were examined to ensure accuracy. The outliers are caused by the largest bag type (Lunch = “L”) as 3 different types were used in 2025. 2024 had only one size bag allowing for a tighter distribution.
No action required. The average tare of each bag type will be applied during summarization/wrangling of data.
## # A tibble: 6 × 5
## year transect quadrat GrossMass BagType
## <dbl> <chr> <dbl> <dbl> <chr>
## 1 2025 102B 3 43.0 L
## 2 2025 102B 4 37.7 L
## 3 2025 101B 5 38.3 L
## 4 2025 102B 1 34.2 L
## 5 2025 103B 4 35.4 L
## 6 2025 101B 4 34.9 LBelow is the comparative visualization of gross mass distributions for 2024 and 2025
Plant Composition
From the 2540 observations from 6 variables, there were no missing or duplicate entries present.
No further action required
## # A tibble: 1 × 6
## Missing_year Missing_transect Missing_quadrat Missing_cover Missing_code
## <int> <int> <int> <int> <int>
## 1 0 0 0 0 0
## # ℹ 1 more variable: Duplicates <int>## # A tibble: 0 × 6
## # Groups: year, transect, quadrat, code [0]
## # ℹ 6 variables: year <dbl>, date <dttm>, transect <chr>, quadrat <dbl>,
## # code <chr>, cover <dbl>Below are comparative summaries over coverage on plant composition.
## # A tibble: 2 × 4
## year Mean SD SE
## <dbl> <dbl> <dbl> <dbl>
## 1 2024 12.7 21.6 0.620
## 2 2025 14.7 23.6 0.647Initial run of detecting outliers was difficult to evaluate as Daubenmire values of 63 and above were triggered as an outlier giving too many results to be credible.
Action taken: group the outliers by code to discover single anomalies such as unlikely plant codes or misinterpreted values and investigate those through standard procedures (i.e. original data sheet, hard copies).
Result: No incorrect entries of data were found.
##
## ANPA ARCA ARPU BAREG CAREX CRUST DEPI HERBLIT JUHO JUNI
## 1 1 2 43 11 33 1 43 2 1
## JUSC MEOF MUSP PASM POTA ROSA SDHERB SWODE SYAL
## 8 1 1 4 3 1 2 49 1Below shows the distribution of plant coverage (note Daubemire hold
specific midpoint values) for 2024 and 2025.
Belt Transects
Reorganization of the data was necessary for ensuring the quality of data output. This is due to the fact the HeightHit entries were populated by two separate conditions thus a different meaning in values 1)“Woody plants ≤ 1 m in height: record species and height in cm” 2)“Woody plants ≥ 1 m in height: record species and start and end points along transect”
For the purposes of examining whether there were any missing/duplicate entries, the two conditions were separated and examined individually.
Woody plants measurements less then 1 m (height in cm) found no missing entries but an abundance of duplicated values. This is due to the nature of the sampling itself as each transect and species are being repeated for each observation (ex: there may be 10 ARTR measured at 20 cm on the same transect, as shown in the tibble below)
No further action required.
## # A tibble: 10 × 4
## year transect species Height_cm
## <dbl> <chr> <chr> <dbl>
## 1 2024 49 ARCA 5
## 2 2024 49 ARCA 8
## 3 2024 47 ARFR 20
## 4 2024 54 ARTR 26
## 5 2024 54 ARTR 20
## 6 2024 38 ARTR 20
## 7 2024 48 ARTR 20
## 8 2024 48 ARTR 13
## 9 2024 39 ARTR 19
## 10 2024 51 ARTR 35When checking for outliers here too shown an abundance that fell out the mainly upper bounds of 2 standard deviations.
Actions taken: Again the outliers were grouped to their specified species. Any anomalies or misinterpereted values were checked over.
##
## ARCA ARTR JUSC RHAR RHTR RIUV ROSA SYAL
## 11 10 24 1 13 6 8 6Below is a visualization of woody plant height distribution for both 2024 and 2025
Woody plant measurement exceeding 1 m (measuring point to point lengths) were also examined for missing/duplicated measurements.
Result: No missing entries detected. While there were some duplicate outcomes in terms of length measurements along the transect, there were no corresponding duplicates when it came to stop/end points.
No further action required.
## # A tibble: 1 × 5
## Missing_year Missing_transect Missing_species Missing_length Duplicates
## <int> <int> <int> <int> <int>
## 1 0 0 0 0 8Below displays the distributions for the canopy coverages for 2024 and 2025
Assessment and Recommended Mitigations:
Finalized statment:
The data summarized above included few errors (mostly due to duplicated entries). These entries have been corrected and noted for future use. Actions to verify, correct or modify data were conducted by R script manipulations, physical examination of the sample, review of excel spreadsheets and inspection of hard copy materials.
Recommended mitigations:
• Improved categorization of data – information put on to a spreadsheet should have uniformity in the sense that all entries are following the conditions of the data desired. (Case: Belt Transect entries should have had two columns for each data type; i.e. “Height measured/≤1m” and “start-end points/≥1m” should be recorded as different types of data being collected.
• Universal bag sizes can reduce the number of outliers detected and will simplify tare calculations.
Ongoing work:
• Summarize data/data wrangling
• Update/modify/clarify species list to ensure accuracy
• Protocol to methods drafting